The prognosis of patients with relapsed acute myeloid leukemia (AML) harboring FLT3 mutations is extremely poor. We have demonstrated that microenvironment/leukemia interactions play a major role in the chemoresistance of leukemic stem cells and that the SDF-1a/CXCR4 axis is a key regulator of this interaction. We recently discovered that Sorafenib is a superb FLT3-ITD inhibitor with high clinical single agent activity, including complete remissions in Phase I studies 1,2. High CXCR4 levels are associated with poor prognosis4, and FLT3 mutations upregulate CXCR45. We recently reported that inhibition of CXCR4 with an analogue of the FDA approved CXCR4 inhibitor AMD3100( Plerixafar) resulted in mobilization of leukemic cells and sensitization to the pro-apoptotic effects of Sorafenib6. G- CSF cleaves SDF-1, down-regulates CXCR4, was found to be highly synergistic with AMD3100 in mobilization hematopoietic stem cells9 and was recently approved by FDA. We have reported that AML patients in remission, who were treated with AMD3100/G-CSF, had massive egress of AML cells into the circulation, providing first proof of principle for leukemia cell mobilization10. In addition, we observed preferential mobilization of AML over normal cells, both in AML patients in CR and in patients with active disease who received AMD3100/G-CSF as part of a preparative regimen followed by SCT, further supporting the validity of this therapeutic concept. Of note, Sorafenib has no toxicity against normal hematopoietic cells. Based on these findings, we propose to test the hypothesis that mobilization of leukemic stem cells by disruption of SDF-1a/CXCR4 by AMD3100/G-CSF results in improved anti-leukemia activity of Sorafenib in AML patients with mutant FLT3. In addition, we will study the effects of this targeted therapy on non-mobilized AML blasts. We will conduct a clinical trial to determine the safety and efficacy of AMD3100, G-CSF and escalating doses of the FLT3-ITD inhibitor Sorafenib in patients with AML harboring FLT3-ITD mutations and study the in vivo biological effects of disrupting SDF-1a/CXCR4 interactions by AMD3100/G-CSF. Patients with FLT3-ITD mutations can have variable numbers of cells with three distinct FLT3 genotypes : heterozygous FLT3-ITD, homozygous FLT3-ITD and wild type FLT3. Therefore, we will also determine, by single-cell PCR, the effects of this therapy against cells carrying these genotypes, during mobilization with AMD3100/G-CSF and following treatment with sorafenib. Taken together, these studies will establish the safety and anti-leukemia activity of AMD3100/G- CSF/sorafenib and provide a comprehensive assessment of mechanisms involved in leukemia cell mobilization and inhibition of cell signaling.